To DVD or not to DVD?

[orionhunter34]

What is the difference between recording in DVD compliant format quality or the standard mpeg2 format quality at the same bit rate. I created some custom DVD formats and custom mpeg2 (non-dvd) formats and it appears to me the mpeg2 format looks better than DVD format but are they both not mpeg2?

[stanger89]

Basically the ability to directly make a DVD-Video out of it, instead of having to transcode it. If you aren't making DVDs, then there's not really any tangible benefit to using one or the other.

[korben_dallas]

I essentially asked the same question here http://www.sagetv.com/forums/showthread.php?t=9925 just a day or two ago, but no one commented.

The only appreciable difference that I can see is that all the DVD prefixed qualities are Variable Bit Rate.

I had some problems setting up 2.2.4, but finally got it working today, and I set my default recording quality to Best.

Once I get some programs recorded, I'll throw them into NeroVision and see if it wants to re-encode/transcode.

I do remember someone saying the outputstreamtype must be equal to 10, which Best quality is, so that is where I am starting my testing.

[orionhunter34]

My experience with NeroVision is that it does not re-encode the mpeg2 format files as long as the bit rate is below a certain level (say for example 9000000 though I'm not sure this is the exact number) when making a DVD. It will not also re-encode the file when just exporting it to mpeg, for say editing commercials as long as you enable smart encoding, but you must enable it more than once. Not just in the dvd screen menu but also again in the export screen menu. However, I have been trying mpeg-vcr for editing commercials and it is way better than Nero. It takes less than 5 minutes for it to create the new file, about the same amount of time as it would take for windows to copy it. It takes Nero at least twice as long. Also with mpeg-vcr (unlike Nero) you can edit 12000000 bit (MPEG-MAX) files with no re-encoding.

[dvd_maniac]

Does anybody know the technical aspects of the recording qualities?

Maybe the reason that one is better than the other at the same bitrate is that one is interlaced and the other is progressive. I am not sure of the exact technical differences however I read that one looks better on a digital display and the other looks better on an analog display.

[stanger89]

Nope, unless you insert the Inverse Telecine flag (which AFIAK, nothing uses) it's all interlaced. My understanding is that it's just different packaging of the data. The quality difference is probably due to the DVD streams being VBR vs CBR for the rest of the settings. VBR can give you a little better quality at a given file size.

[orionhunter34]

Originally Posted by korben_dallas

I essentially asked the same question here http://www.sagetv.com/forums/showthread.php?t=9925 just a day or two ago, but no one commented.

The only appreciable difference that I can see is that all the DVD prefixed qualities are Variable Bit Rate.

Okay here's my theory about variable bit rate (VBR). VBR is good because it can produce smaller sized files then FBR (fixed bit rate) and still produce the same quality output. Here is how I guess it works. VBR will only encode the amount of bits per video frame that it thinks it needs while FBR assigns a prescribed amount for each frame. Where VBR shines is that it will use less bits for some frames and more for others on an as need basis without compromising quality. If the average amount of bits used is generally less than FBR then a smaller file size is produced but the video quality is the same or better.

Example:
solution 1: VBR - 10Million bps to 12Million bps.
solution 2: FBR - 11Million bps.

solution 1 and 2 will produce the same size file if the VBR rate average is the same as the FBR rate. Also, if a frame required more than 11Million bps it was able to get it only from VBR.
The downside with VBR is you are uncertain what size the file will be since you do not know what the average rate will be for each video file. (trial & error)
I heard that most tv productions are done using FBR. And that VBR was done later for file efficiency. I would think FBR would be easier to implement but since it is all hardware encoded it probably doesn't matter as far as your cpu is concerned.

[korben_dallas]

It depends if the VBR encoder has bit reservoir calculations. If it does, then it can take the saved bits from low data sequences and use them for higher data sequences. Of course, this produces a more complex stream to decode, which in turn requires the end user to have more powerful (expensive) decoding hardware/software.